Background In industrialised countries, approximately 5–20% of women smoke during pregnancy. We aim to study the association between smoking during pregnancy and adverse perinatal outcomes.
Methods A retrospective population–based cohort study using data on all singleton births between 1991 and 2010 (n=1 164 953) derived from the Finnish Medical Birth Register.
Results Of all the mothers included, 82.3% were non-smokers, 2.6% quit smoking during the first trimester of pregnancy, 12.5% smoked throughout pregnancy and 2.7% had no information on smoking. Continuing smoking after the first trimester of the pregnancy was associated with an increased prevalence of admission to a neonatal intensive care unit, stillbirth, preterm birth (<37 gestational weeks), low birth weight (LBW, <2500 g), small for gestational age (SGA, <−2 SDs) and major congenital anomaly compared with non-smokers. Smoking cessation reduced the risk of prematurity, stillbirth, LBW and SGA close to or at similar levels as those of non-smokers. Tobacco exposure in early pregnancy resulted in a 19% increased prevalence of admission to neonatal intensive care unit and a 22% increased prevalence of major congenital anomaly compared with non-smokers.
Conclusions Smoking cessation appeared to reduce pregnancy risks close to those of non-smoking peers. Exposure to early pregnancy smoking was, however, associated with an increased admission to neonatal intensive care and an increased prevalence of major congenital anomalies.
Statistics from Altmetric.com
Socioeconomic gradient is negatively associated with the burden of adverse perinatal outcomes1 since adverse health behaviour and lifestyle factors vary by socioeconomic status (SES).1 ,2 For example, women with a low SES are more likely to be smokers3 which may partly explain the association between SES and adverse perinatal outcomes.4–6 An association between smoking and several adverse perinatal outcomes, such as small for gestational age (SGA),3 ,7–9 low birth weight (LBW),3 ,10 perinatal death9 ,11 and prematurity7–10 is well established in several large population-based studies. Smoking has also been shown to be strongly associated with children's neonatal12 and long-term morbidity.13 ,14 However, previous sibling studies have shown that association between tobacco exposure during pregnancy and long-term outcomes, such as externalising behaviours and education was modified by environmental factors, such as SES and genetic factors.15 ,16
In 2010, in European Union member states, smoking rates during pregnancy reportedly varied from 5% to 19%.17 Previous population-based studies showed that smoking cessation in early pregnancy during the first trimester of pregnancy decreased the risk of SGA,8 ,9 stillbirth11 and preterm birth8 close to or at similar levels as non-smokers. Further, a systematic review reported smoking cessation to be associated with an increased risk of several birth defects.18 However, an association between smoking cessation during pregnancy and some adverse outcomes such as congenital anomalies and admission to neonatal intensive care unit were less studied in previous population-based studies. The present study was aimed to evaluate associations between the smoking habit during pregnancy (non-smokers, women who gave up smoking during the first trimester of the pregnancy, smokers and missing information on smoking) and several perinatal outcomes (admission to a neonatal unit, stillbirth, prematurity, LBW, SGA and major congenital anomaly) using a large 20-year population-based data gathered from the Finnish Medical Birth Register (MBR). Data included all singleton births (n=1 164 953) born from 1991 to 2010 in Finland, a country with around 5.5 million residents and publicly funded delivery care services.
Data and population
The source of the data was the Finnish MBR, which is currently maintained by the National Institute for Health and Welfare (THL) that authorised the use of health register data as required by national data protection legislation. Informed consent of the registered individuals was not needed since we used anonymised data and none of the study subjects were contacted personally.
The MBR contains sociodemographics, pregnancy characteristics, delivery characteristics and perinatal outcomes of all live births or stillbirths delivered after the 22nd gestational week, or weighing 500 g or more. The data included all singleton births (n=1 164 953) that took place in Finland during 1991–2010. Multiple pregnancies (n=36 078) were excluded since they carry a higher risk for complications.
Variables and definitions
Maternal smoking categories were based on self-reported data, which were recorded as non-smoking, quit smoking during the first trimester, or smoking after the first trimester, that is, persistent smokers. Information on smoking was missing for 2.7% (n=30 930) of cases, that was included in analyses as a separate group. SES groups were based on maternal occupation at the time of delivery. Participants were categorised as upper white-collar worker, such as lawyer and teacher, lower white-collar worker, such as nurse and secretary, blue-collar worker, such as waiter and cooker, or other, based on Finland’s national classification of occupations19 based on the international recommendations. The ‘other’ category included entrepreneurs, students, retired, unemployed, housewives and unclassifiable occupations. Information on SES was self-reported and was missing for 13.3% (n=154 575) of cases, which were therefore included in analyses as a separate group.
Estimation of gestational age was based on recorded last menstrual period, unless there was a discrepancy of more than 7 or 14 days at the first trimester or second trimester ultrasonography measurements, respectively. Preterm birth was defined as <37 gestational weeks. Parity was recorded as either nulliparous (no previous births) or multiparous (at least one previous birth) based on MBR information on previous births. The newborn was considered SGA if the birthweight was more than two SDs below the sex-specific and parity-specific mean weight for the gestational age using the Finnish population-based birth curves from 1996 to 200820 as suggested by the International Societies of Pediatric Endocrinology and the Growth Hormone Research Society,21 and LBW was defined as a birth weight of less than 2500 g. Information on in-vitro fertilisation (IVF) included intracellular sperm injection (ICSI) and frozen embryo transfers (FET). Maternal anaemia was defined as a haemoglobin level below or equal to 6.2 mmol/L. Marital status was recorded as married/living with a partner or single. Mode of delivery was spontaneous vaginal, breech, forceps, vacuum assisted or caesarean section. The study period 1991–2010 was divided into four time periods (1991–1995, 1996–2000, 2001–2005 and 2006–2010) to investigate secular trends. Diagnoses of newborns admitted to a neonatal unit were defined using International Classification of Diseases (ICD)-9 and -10 codes.
Information on major congenital anomalies among stillborn or liveborn children was obtained from the Finnish National Congenital Malformations Register established in 1963 and currently maintained by the National Institute for Health and Welfare. The register contains data on major congenital anomalies (yes or no) until 1 year of age gathered from multiple data sources. MBR and Congenital Malformations Register were linked together using encrypted unique personal identification numbers.
Bivariable analyses were performed to evaluate statistical differences in ICD-10 diagnoses between smoking status groups using the χ2 test. Multivariable logistic regression analysis was performed to calculate ORs with 95% CIs between smoking habit and each adverse perinatal outcome. An association between smoking habit and each adverse perinatal outcome was stratified by maternal age, parity, sex and SES. Further, an association between smoking habit and admission to neonatal intensive care unit was adjusted for preeclampsia and gestational diabetes. We performed sensitivity analyses to avoid possible bias arising from missing information on SES; all cases with missing observations were assigned to occupations category values one by one to assess robustness of results under the extreme situation of complete concentration of missingness in specific categories. The data were analysed using SPSS for Windows V.19-0, Chicago, Illinois, USA.
Of all the women with singleton births (n=1 164 953), 82.3% (n=958 177) were non-smokers, 2.6% (n=30 136) quit smoking during the first trimester of pregnancy, 12.5% (n=145 710) smoked during and after the first trimester of pregnancy (table 1). Altogether, 21.9% blue-collar workers smoked, compared with only 3.4% of upper white-collar workers. Women who continued smoking after the first trimester of pregnancy were relatively younger compared with non-smokers, had more frequently experienced prior terminations of pregnancy and were less frequently affected by preeclampsia. The number of women who quit smoking ranged from 1.0% among all upper white-collar workers to 3.6% among all blue-collar workers. Among the smokers, the percentages of women quitting smoking during the first trimester were 24% and 14%, respectively. Women who quit smoking tended to be nulliparous and relatively young, gave birth at later gestational weeks, and had an infant with a higher birth weight in comparison with women who continued smoking after the first trimester of pregnancy. Non-smokers were significantly more likely to be married or living with a partner compared with the other groups. Prevalence of each adverse perinatal outcome according to SES and smoking status groups are presented in table 2.
Table 3 presents stratified associations between smoking habit during pregnancy (the reference group being non-smokers) and each adverse perinatal outcome. Continuing smoking after the first trimester of pregnancy resulted in an increased prevalence of each adverse perinatal outcomes than in non-smokers; admission to neonatal intensive care unit (adjusted OR (aOR) 1.27, 95% CI 1.25 to 1.30), stillbirth (aOR 1.13, 95% CI 1.06 to 1.20), preterm birth (aOR 1.39, 95% CI 1.36 to 1.43), LBW (aOR 2.02, 95% CI 1.97 to 2.07), SGA (aOR 2.47, 95% CI 2.41 to 2.53) and major congenital anomalies (aOR 1.13, 95% CI 1.09 to 1.16). Quitting smoking during the first trimester of pregnancy resulted in an equal prevalence of stillbirth and preterm birth with non-smokers, and prevalence of LBW and SGA close to non-smokers. Women who smoked early in pregnancy but quit in the first trimester still had 19% (aOR 1.19, 95% CI 1.15 to 1.24) elevated risk for admission to a neonatal intensive care unit and a 22% (aOR 1.22, 95% CI 1.15 to 1.30) increased prevalence of major congenital anomalies. Sensitivity analyses assigning observations with missing information on SES to each SES category in turn did not meaningfully alter our findings (data not shown).
Among infants admitted to a neonatal unit, reasons for admission differed significantly across smoking status groups (table 4). There were significant differences between smoking status groups in the frequencies of respiratory distress, asphyxia, neonatal infection and hypoglycaemia among infants admitted to intensive neonatal care.
The aim of the present study was to evaluate the association between smoking behaviour during pregnancy and adverse perinatal outcomes. In Finland, smoking throughout the pregnancy among singleton births decreased over the 20-year period 1991–2010 (range from 14.2% to 10.3%), and the total prevalence (12.5%) was comparable to that in other European countries in 2010.17 The prevalence of smoking cessation during the first trimester and persistent smoking varied substantially with SES, as found previously.2 ,4 ,22 The most important finding of the present study was that smoking exposure during early pregnancy, even with first trimester cessation, was associated with an increased admission to intensive care and major congenital anomalies compared with non-smokers. Continuing smoking after the first trimester of pregnancy was identified as a risk factor of stillbirth, LBW, SGA, prematurity and major congenital anomaly that was in line with previous evidence.3 ,7–11 Quitting smoking during the first trimester of pregnancy reduced prevalence of stillbirth, LBW, SGA and preterm birth equal or close to those of non-smokers that was also in line with previous studies.8 ,9 ,11
The present study has several strengths. The most important one is the high-quality data gathered from the mandatory, national Finnish MBR including the entire population. To the best of our knowledge, the present study population is one of the largest studies to date. Self-reported smoking behaviour during pregnancy was indicated to be a valid way to study tobacco exposure during pregnancy in observational studies compared with biomarkers such cotinine measurements that are not repeated.23 Smoking habits during pregnancy are relatively well covered in the Finnish MBR,24 and information on smoking was collected by using the same classification during for the whole study period. A change in smoking behaviour during pregnancy was assessed in antenatal visits that is likely to be more reliable than assessment at delivery time only.
A possible limitation was that smoking has been shown to be a dose-dependent risk factor,3 ,25 but we did not have sufficient data to examine a possible dose response. Another possible limitation was that SES was determined solely on the basis of maternal occupation and we had no information on the father's occupational status or family income. However, occupation is related to education and income in Finland, and is an available, appropriate indicator on socioeconomic health differences. 1 ,26 ,27 Another possible limitation was that the categorisation of parturients’ SES is often challenging since they were typically young, without occupation, such as students, or stayed at home to take care or their children. Information on SES is self-reported and some hospitals consider it optional: therefore, the information was missing for 13.3% of the cases which might have biased our results. However, sensitivity analyses in which we assumed the extreme situation of all observations with missing SES values arising from one occupational group did not affect our results. This provides some evidence that missing data on SES were not associated with the exposure-outcome association of interest.
The association between tobacco exposure associated with early pregnancy (eg, in the setting of first trimester smoking cessation) with congenital anomalies and neonatal intensive care admission has not been well described previously. We demonstrated that antenatal tobacco exposure in early pregnancy was associated with a 19% increased admission to a neonatal intensive care unit and a 22% increased prevalence of major congenital anomalies. One speculative, possible explanation for an increased admission to a neonatal intensive care unit among infants of women who quit smoking would be the nicotine withdrawal effect previously shown in laboratory animal studies; infants exposed to smoking in early pregnancy were first exposed to nicotine and then to nicotine withdrawal, as found within the experimental setting using rats.28 Maternal nicotine exposure during gestation and lactation was found to have no effect on the lung parenchyma of offspring up to weaning, but deterioration and other structural changes were apparent by around six postnatal weeks and continued until 12 weeks. These results together with the known longer half-life (3–4-fold) of nicotine29 and higher levels of nicotine in newborns’ fetal circulation and amniotic fluid in comparison with adults suggest that adverse effects of nicotine on cell integrity occur after a substantial delay and only after nicotine withdrawal, rather than during exposure. Results of the present study were in accordance with that since infants of mothers who quit smoking in early pregnancy were admitted to neonatal care slightly more often due to asphyxia and respiratory distress compared to infants of non-smokers. Our results were also in line with previous studies that reported a decreased risk of infant respiratory distress syndrome in infants of smokers compared to infants of non-smokers.30 ,31
The association between smoking exposure and major congenital anomalies was in agreement with the results of recent systematic review that suggested smoking during pregnancy to be associated with several birth defects.18 However, our results showed that even tobacco exposure restricted to early pregnancy but during organ development was associated with an increased risk of major congenital anomalies compared to non-smokers.
Our results show that smoking and quitting smoking during pregnancy have clear socioeconomic gradients. Early smoking cessation seems to reduce pregnancy risks of stillbirth, preterm birth, SGA and LBW close to those of non-smokers, but not risks of major congenital anomaly and need for neonatal care. Infants who were exposed to tobacco in early pregnancy only may be at higher risk during the neonatal period compared to unexposed infants, and it seems clear that women who want to become pregnant should give up smoking before pregnancy. Further studies are needed to replicate our findings and to unravel the mechanisms behind these clinical findings.
What is already known on this subject
Smoking during pregnancy is a socially patterned dose-dependent risk factor for several adverse perinatal outcomes, such as small for gestational age and prematurity.
What this study adds
Exposure to tobacco during early pregnancy was associated with an increased admission for neonatal intensive care and an increased prevalence of major congenital anomalies. However, early smoking cessation seemed to reduce the risks of preterm birth, small for gestational age and low birth weight similar or close to babies of non-smokers.
Health education about consequences of cigarette smoking during pregnancy should be offered for young women planning pregnancy.
Contributors SR designed the study, managed the dataset, performed statistical analyses, and wrote the first draft of the paper, revised the paper and approved the final manuscript as suggested. US designed the study, was statistical advisor, wrote the paper and approved the final manuscript as suggested. MG designed the study, was statistical advisor, wrote and revised the paper and approved the final manuscript as suggested. MRK designed the study, was statistical advisor, wrote and revised the paper and approved the final manuscript as suggested. TH-V designed the study, revised the paper and approved the final manuscript as suggested. JS designed the study, revised the paper and approved the final manuscript as suggested. SH designed the study, was statistical advisor, wrote and revised the paper and approved the final manuscript as suggested.
Competing interests None.
Ethics approval Permission to use the confidential register data in this study was accorded on 16th February 2012 by the National Institute for Health and Welfare (THL) in Finland (Reference number 1749/5.05.00/2011).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.